Surveying biomedical relation extraction: a critical examination of current datasets and the proposal of a new resource.

Natural language processing (NLP) has become an essential technique in various fields, offering a wide range of possibilities for analyzing data and developing diverse NLP tasks. In the biomedical domain, understanding the complex relationships between compounds and proteins is critical, especially in the context of signal transduction and biochemical pathways. Among these relationships, protein-protein interactions (PPIs) are of particular interest, given their potential to trigger a variety of biological reactions. To improve the ability to predict PPI events, we propose the protein event detection dataset (PEDD), which comprises 6823 abstracts, 39 488 sentences and 182 937 gene pairs. Our PEDD dataset has been utilized in the AI CUP Biomedical Paper Analysis competition, where systems are challenged to predict 12 different relation types. In this paper, we review the state-of-the-art relation extraction research and provide an overview of the PEDD's compilation process. Furthermore, we present the results of the PPI extraction competition and evaluate several language models' performances on the PEDD. This paper's outcomes will provide a valuable roadmap for future studies on protein event detection in NLP. By addressing this critical challenge, we hope to enable breakthroughs in drug discovery and enhance our understanding of the molecular mechanisms underlying various diseases.

Case 329.

HISTORY: A 45-year-old female patient who was previously healthy presented after several weeks of fullness in the right upper quadrant of the abdomen. The patient did not experience pain, nausea, vomiting, or jaundice, and had no contributory past medical or surgical history, including no history of malignancy. Upon examination, vital signs were within normal limits and the patient was appeared well with soft palpable fullness in the right upper quadrant. The abdomen was nontender and nondistended. Laboratory investigation revealed no abnormalities, with a normal complete blood cell count and normal serum tumor markers that included α-fetoprotein (<2.0 ng/mL; reference, <8.3 ng/mL), cancer antigen 19-9 (21.6 U/mL; reference, <35 U/mL), and carcinoembryonic antigen (1.3 ng/mL; reference, <5 ng/mL). CT of the abdomen and pelvis was performed with intravenous contrast material in the emergency department (Fig 1). Subsequently, combined MRI and MR cholangiopancreatography of the abdomen was performed with and without intravenous contrast material for further evaluation (Fig 2). CT of the chest performed during the same encounter was unremarkable.

CT and MR Enterography in the Evaluation of Celiac Disease.

Celiac disease is a common inflammatory disease of the small bowel that induces mucosal intestinal lesions. The disease is mediated by an immune response and triggered by the ingestion of gluten in genetically predisposed individuals. Gluten contains gliadin, a component found mostly in wheat, barley, and rye. This process leads to gastrointestinal malabsorption with symptoms such as diarrhea, constipation, abdominal pain, and distention. It has a prevalence of 1%-2% in the general adult population, who present with symptoms at any age, but is more frequently found in adult women in the 3rd or 4th decade of life. Recognition of the disease has increased, but it remains a challenge to diagnose. CT and MR enterography are noninvasive studies used for evaluation of small bowel neoplasms and inflammatory small bowel pathologic conditions such as celiac disease. The authors review the spectrum of intestinal and extraintestinal findings of celiac disease at CT and MR enterography, as well as its complications, and the importance of recognizing certain imaging features that help in the diagnosis of celiac disease. More common and specific findings of celiac disease such as inversion of the jejunoileal fold pattern and mesenteric lymphadenopathy are reviewed. More uncommon entities that are more frequently associated with refractory or untreated celiac disease, such as ulcerative jejunoileitis, cavitary mesenteric lymph node syndrome, and malignancies including small bowel adenocarcinoma and lymphoma, are described. ©RSNA, 2024 Test Your Knowledge questions for this article are available in the supplemental material. The slide presentation from the RSNA Annual Meeting is available for this article.

Episodic Memory Loss: When Alzheimer's Disease Is Not the Answer.

A 60-year-old man presented to a Neurology Clinic specialized in cognitive disorders to evaluate memory complaints. A comprehensive neuropsychological examination detected an isolated and severe hippocampal memory deficit. Laboratory tests, brain magnetic resonance imaging (MRI), and cerebrospinal fluid (CSF) tests, including Alzheimer's disease (AD) biomarkers, did not show remarkable results. Due to family history of cognitive impairment, we extended the study to non-Alzheimer monogenic mutations (Next Generation Sequencing) detecting a pathogenic variant of the progranulin (PGRN) gene (c.1414-1 G > T) which has been previously associated with the same phenotype. These results should be considered in patients with an Alzheimer-like presentation, negative AD biomarkers' results, and family history of dementia.

Lemann Index for assessing bowel damage in Crohn's disease: a real world study.

BACKGROUND & AIMS: The Lemann Index (LI), an endpoint to measure cumulative structural bowel damage in Crohn's disease (CD), has been recently updated and validated. We applied this to investigate predictors of bowel damage in a real-world cohort. METHODS: We performed a retrospective study (2008-2022) involving two tertiary referral IBD centers in the US. MR or CT enterographies were reviewed by study radiologists and endoscopy reports by study gastroenterologists, to calculate LI. Baseline and follow-up LI were calculated. We defined high bowel damage as LI ≥2. Factors associated with high LI were identified in patients with ≥2 LI scores using multivariate logistic regression and then assessed for a change in LI (increase vs. no change/decrease) using a multivariate linear mixed-effects model. RESULTS: 447 patients with CD had a median first LI of 7 [IQR, 1.25-14.55]. Median LI scores were significantly different when categorized by disease duration; 2.0 [IQR, 0.6-5.9] for <2 years, 2.6 [IQR, 0.6-9.6] for ≥2 and <10 years, and 12.5 [IQR, 6.4-21.5] for ≥10 years with a p <0.01. Disease duration, presence of perianal disease, elevated C-reactive protein, and Harvey-Bradshaw index, were associated with a high LI at inclusion and increase in LI during follow-up (all p <0.01). CONCLUSIONS: The updated LI quantified cross-sectional and longitudinal cumulative bowel damage in a real-world cohort of patients with CD with predictors identified for a longitudinal increase in LI. Further studies for prospective validation of LI and identification of multi-omic predictors of bowel damage are needed.

Alterations in Lysosomal, Glial and Neurodegenerative Biomarkers in Patients with Sporadic and Genetic Forms of Frontotemporal Dementia.

Background: Frontotemporal dementia (FTD) is the most common cause of early-onset dementia with 10-20% of cases caused by mutations in one of three genes: GRN, C9orf72, or MAPT. To effectively develop therapeutics for FTD, the identification and characterization of biomarkers to understand disease pathogenesis and evaluate the impact of specific therapeutic strategies on the target biology as well as the underlying disease pathology are essential. Moreover, tracking the longitudinal changes of these biomarkers throughout disease progression is crucial to discern their correlation with clinical manifestations for potential prognostic usage. Methods: We conducted a comprehensive investigation of biomarkers indicative of lysosomal biology, glial cell activation, synaptic and neuronal health in cerebrospinal fluid (CSF) and plasma from non-carrier controls, sporadic FTD (symptomatic non-carriers) and symptomatic carriers of mutations in GRN, C9orf72, or MAPT, as well as asymptomatic GRN mutation carriers. We also assessed the longitudinal changes of biomarkers in GRN mutation carriers. Furthermore, we examined biomarker levels in disease impacted brain regions including middle temporal gyrus (MTG) and superior frontal gyrus (SFG) and disease-unaffected inferior occipital gyrus (IOG) from sporadic FTD and symptomatic GRN carriers. Results: We confirmed glucosylsphingosine (GlcSph), a lysosomal biomarker regulated by progranulin, was elevated in the plasma from GRN mutation carriers, both symptomatic and asymptomatic. GlcSph and other lysosomal biomarkers such as ganglioside GM2 and globoside GB3 were increased in the disease affected SFG and MTG regions from sporadic FTD and symptomatic GRN mutation carriers, but not in the IOG, compared to the same brain regions from controls. The glial biomarkers GFAP in plasma and YKL40 in CSF were elevated in asymptomatic GRN carriers, and all symptomatic groups, except the symptomatic C9orf72 mutation group. YKL40 was also increased in SFG and MTG regions from sporadic FTD and symptomatic GRN mutation carriers. Neuronal injury and degeneration biomarkers NfL in CSF and plasma, and UCHL1 in CSF were elevated in patients with all forms of FTD. Synaptic biomarkers NPTXR, NPTX1/2, and VGF were reduced in CSF from patients with all forms of FTD, with the most pronounced reductions observed in symptomatic MAPT mutation carriers. Furthermore, we demonstrated plasma NfL was significantly positively correlated with disease severity as measured by CDR+NACC FTLD SB in genetic forms of FTD and CSF NPTXR was significantly negatively correlated with CDR+NACC FTLD SB in symptomatic GRN and MAPT mutation carriers. Conclusions: In conclusion, our comprehensive investigation replicated alterations in biofluid biomarkers indicative of lysosomal function, glial activation, synaptic and neuronal health across sporadic and genetic forms of FTD and unveiled novel insights into the dysregulation of these biomarkers within brain tissues from patients with GRN mutations. The observed correlations between biomarkers and disease severity open promising avenues for prognostic applications and for indicators of drug efficacy in clinical trials. Our data also implicated a complicated relationship between biofluid and tissue biomarker changes and future investigations should delve into the mechanistic underpinnings of these biomarkers, which will serve as a foundation for the development of targeted therapeutics for FTD.

The overview of the BioRED (Biomedical Relation Extraction Dataset) track at BioCreative VIII.

The BioRED track at BioCreative VIII calls for a community effort to identify, semantically categorize, and highlight the novelty factor of the relationships between biomedical entities in unstructured text. Relation extraction is crucial for many biomedical natural language processing (NLP) applications, from drug discovery to custom medical solutions. The BioRED track simulates a real-world application of biomedical relationship extraction, and as such, considers multiple biomedical entity types, normalized to their specific corresponding database identifiers, as well as defines relationships between them in the documents. The challenge consisted of two subtasks: (i) in Subtask 1, participants were given the article text and human expert annotated entities, and were asked to extract the relation pairs, identify their semantic type and the novelty factor, and (ii) in Subtask 2, participants were given only the article text, and were asked to build an end-to-end system that could identify and categorize the relationships and their novelty. We received a total of 94 submissions from 14 teams worldwide. The highest F-score performances achieved for the Subtask 1 were: 77.17% for relation pair identification, 58.95% for relation type identification, 59.22% for novelty identification, and 44.55% when evaluating all of the above aspects of the comprehensive relation extraction. The highest F-score performances achieved for the Subtask 2 were: 55.84% for relation pair, 43.03% for relation type, 42.74% for novelty, and 32.75% for comprehensive relation extraction. The entire BioRED track dataset and other challenge materials are available at https://ftp.ncbi.nlm.nih.gov/pub/lu/BC8-BioRED-track/ and https://codalab.lisn.upsaclay.fr/competitions/13377 and https://codalab.lisn.upsaclay.fr/competitions/13378. Database URL: https://ftp.ncbi.nlm.nih.gov/pub/lu/BC8-BioRED-track/https://codalab.lisn.upsaclay.fr/competitions/13377https://codalab.lisn.upsaclay.fr/competitions/13378.